A Novel Personalized Recommendation Model for Computing Advertising Based on User Acceptance Evaluation

Xie, Yanqiu; Huang, Yanli

doi:10.1109/access.2023.3339839

Cited by 2 publications

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Xie and Huang [13] introduced a personalized recommendation model for computing advertising based on user acceptance evaluation, emphasizing user experience and acceptance in advertising. Qian et al [14] focused on intent disentanglement and feature self-supervision for novel recommendation, aiming to provide more relevant and interpretable recommendations.…”

Section: Review Of Existing Modelsmentioning

confidence: 99%

Sentiment-aware Content Recommendation using LSTM-based Collaborative Filtering

Morzelona,

Batra

2022

RJCSE

View full text Add to dashboard Cite

The need for this work arises from the ever-growing demand for more precise and efficient content recommendation systems. Existing methods, while serving their purpose, are not without limitations. They often struggle to capture the intricate nuances of user sentiment, which is crucial for personalized content recommendations. This limitation results in suboptimal precision, accuracy, recall, and speed, ultimately leading to less satisfying user experiences. To address these shortcomings, this paper introduces a novel approach, leveraging the fusion of Bidirectional Long Short-Term Memory (BiLSTM) with BERT, combined with a Graph Convolutional Network (GCN). This innovative model brings together the power of natural language processing and graph-based techniques, enhancing the efficiency of sentiment-aware content recommendation and filtering. The use of BiLSTM and BERT allows our model to grasp the contextual intricacies of user preferences, making it more adept at discerning sentiment in textual data samples. The incorporation of GCN further enriches our model's capabilities by leveraging the underlying connections between users and content, enabling it to provide more personalized recommendations. The advantages of our proposed approach become evident through rigorous testing on multiple contextual datasets. Compared to existing methods, our model exhibits notable improvements, with an 8.3% boost in precision, an 8.5% increase in accuracy, a 5.9% rise in recall, and a 4.5% enhancement in speed. Moreover, our model achieves a 7.5% better Area Under the Curve (AUC), solidifying its effectiveness in sentiment-aware content recommendation. In conclusion, this work fills a crucial gap in the realm of content recommendation by introducing a sophisticated model that excels in capturing user sentiment and delivering more precise and efficient recommendations. These advancements pave the way for improved user experiences and hold significant implications for a wide range of applications, from e-commerce to personalized content delivery platforms.

show abstract

Section: Review Of Existing Modelsmentioning

confidence: 99%